collectionplan

type Collection struct {
	Name  string
	Files []string
	Bytes int64
	// Folders are the folders of the archive the files come from, relative to its root ("" is the root itself).
	Folders []string
}

type Duplicate struct {
	Path string
	Of   string
}

type Entry struct {
	// Path is slash-separated and relative to the root of the archive.
	Path    string
	Size    int64
	ModTime time.Time
	// Hash is the hash of the content when the scanner computed one; two entries with the same non-empty hash are the same file.
	Hash string
	// Link is set for a symbolic link: it is not followed, and the plan leaves it out.
	Link bool
	// Problem is set when the scanner could not read the entry (a folder or a file it has no right to): it is listed, not lost.
	Problem string
}

type Excluded struct {
	Path   string
	Reason string
}

type Options struct {
	// Readable says whether a file extension (lower case, with its dot) is one the reader can turn into text. Nil means DefaultReadable.
	Readable func(ext string) bool
	// MinFiles is how many readable files a folder needs to be a collection of its own; smaller folders are gathered into one collection
	// named MiscName. Default 2: only a folder of a single file is gathered. A higher value gathers real groups: on the test bench
	// (scripts/collections-eval), with half of the folders right, gathering the folders of up to 1, 2 and 4 files leaves 52 %, 54 % and
	// 58 % of the files to fix afterwards, against 51 % when nothing is gathered.
	MinFiles int
	// MaxFiles is the size above which a collection is split along its subfolders, when it has subfolders big enough to stand alone.
	// Default 400.
	MaxFiles int
	// MaxPlanFiles refuses a listing longer than this (ErrTooManyFiles). Default 100000.
	MaxPlanFiles int
	// MiscName is the name of the collection that gathers the small folders and the files at the root. Default "Miscellaneous".
	MiscName string
}

type Plan struct {
	Collections []Collection
	Excluded    []Excluded
	Duplicates  []Duplicate
}